In this article, a web-based application, that allows creation of datasets and sets of rules to build a fuzzy logic-based decision support system (DSS), is proposed. The web application allows creating and deploying many DSSs that can have different specializations and simultaneously be used by many other systems, including cyber-physical system (CPS), as a mean of decision support, via an application programming interface (API), using HTTP requests. The creation of the DSS, based on the fuzzy logic principle, and filling data for it can be done by both an expert human and an application that can handle and transfer data from a specific work area through the Internet. A configured DSS can be used by the CPS and can receive crisp values as input and provide crisp values as output after processing data by the fuzzy logic system. The main technologies that are used in the construction of the web-based application and its architecture are also considered. Fuzzy logic, web-based application, decision support system, cyber-physical system, fuzzy IntelITSIS'2022: 3rd International Workshop on Intelligent Information Technologies and Systems of Information Security, March 23-25,
necessitate a high level of expertise. information.
Autonomous CPSs are systems that can make decisions and take actions without the need for
human interaction. The challenge of decision-making is serious in such systems, and
many
technological approaches are utilized to attain autonomy, which are based on the architecture of the
CPS and the tasks it should do. For example, applying specialized processors for decision-making
means can be done [
Because of its capacity to successfully tackle a wide range of issues, concerning the decisionmaking process in CPSs, fuzzy systems have been widely used.
One of the most important areas for the usage of fuzzy set theory are fuzzy systems. They employ
fuzzy logic to give a conceptual framework for knowledge representation and reasoning in the context
of imprecision and uncertainty. Due to their ability to incorporate human expert knowledge, handle
imprecision and uncertainty, and describe the behavior of complex systems without requiring a
precise mathematical model, fuzzy systems have been successfully applied to many application fields,
such as control [
Fuzzy logic is similar to the way that human makes a decision. It deals with fuzzy and inaccurate
2022 Copyright for this paper by its authors.
The word fuzzy refers to things that are not clear or indistinct. Any event, process, or function that
is in a state of constant change can’t be defined as true or false all the time and it means that we need
to define such actions in a fuzzy way [
The difference between Boolean logic (Fig. 1) and fuzzy logic (Fig. 2) is that, unlike Boolean logic, fuzzy logic operates differentiated values and they are denoted by a numbers ranging from zero to one, where one means absolute truth and zero represents absolute falsehood. A number indicating a value in fuzzy systems is called the truth value.
Thus, it can be argued that fuzzy logic is not fuzzy in its content, but is such that may be used to describe the fuzziness.
Fuzzy logic apps often use non-numeric values to facilitate the expression of rules and facts by an expert. Linguistic variables, such as temperature, can take on values such as cold and hot. Because natural languages do not always contain enough terms to express a fuzzy scale of meanings, it is common practice to change the meanings of speech using adjectives or adverbs. For example, a linguistic variable can take such values as very cold or very hot.
Linguistic variables, their meanings, and rules are represented in fuzzy logic systems assets - an unordered collection of different elements. It can be written explicitly by enumerating its elements using parentheses.
Fuzzy sets may be taken into consideration as an extension and gross oversimplification of classical units. It may be understood inside the context of set membership. It lets in partial membership because it comprises factors that have various levels of sets withinside the set. We can recognize the distinction between classical set and fuzzy set from this.
The classical set carries factors that fulfill particular properties of membership (Fig. 3), while the fuzzy set carries factors that fulfill obscure properties of membership.
A fuzzy set A˜ in the universe of information U can be defined as a set of ordered pairs and it can be represented mathematically as (1) where one, so that: is the degree of membership of y in
and assumes values in the range from zero to
This fuzziness is well characterized with the aid of using its membership function. In addition, the membership function represents the degree of truth in fuzzy logic (Fig. 4).
The presentation layer includes controllers end endpoints to receive HTTP requests and send HTTP responses. Controllers and endpoints here are used to create, update and delete records in database tables.
Data can be passed to endpoints with the usage of query parameters. Provided API can be used by a human expert or another application and allows the creation of subject areas with sets of necessary rules, terms, and values. Rule’s syntax representation, as well as query parameters’ format, are predefined.
The service layer is responsible for processing data, provided in HTTP requests, and transferring created entities’ objects to the domain layer. Also, it includes fuzzy logic services, that together are the core of fuzzy logic – fuzzy inference system, and can perform operations on fuzzification, defuzzification, and decision-making.
The domain layer represents the business logic and objects’ behavior. It holds application objects and can make base operations on the database.
The proposed relational database is used by the application and includes 4 tables. Its diagram is shown in Figure 6.
FuzzyLogicAreas table record has the AreaName property that corresponds to the subject area that has to be solved. This table can have many records because the application can be used by many CPSs and their subsystems. It has one-to-many relationships with Rules and Terms tables.
Terms table record has a TermName string property that represents a linguistic variable. Terms table stores a set of linguistic variables for the specific subject area. This table has a one-to-many relationship with the Sets table.
Sets table record has Key string property, which is a linguistic value that can be applied to a corresponding linguistic variable, and its numeric value, that is Value double property.
Rule table record has a RuleValue string property, that represents a rule in the specified format, using linguistic variables and their corresponding linguistic values. Rules are defined as an IF-THEN structure.
A fuzzy inference system is the main unit of a fuzzy logic system, which is responsible for doing one main task – making decisions. It uses IF-THEN rules along with basic logic operators.
The fuzzy inference system diagram that is implemented with fuzzy logic services in the web-API application service layer is shown in Figure 7.
Decision-making service operates on rules. Foremost, it defines a set of rules, that are stored in the database and applies it to a set of input and output crisp data.
After that, for each rule, rule strength is established by combining the fuzzified inputs according to rules and the result is a set of rule strength values.
During the rules processing, boolean operators, that are used in the rules’ description are replaced with Zadeh operators, please, check Table 1.
For getting output distribution, all the consequents are combined and passed as a result fuzzy set to the defuzzification service.
Defuzzification service then calculates the arithmetic mean of the result fuzzy set for each output value and converts the arithmetic mean to crisp value by using the output membership function. Finally, a defuzzified, crisp output set is obtained.
Foremost, using the API of the web application, the research subject area was created and filled with data and rules.
Created DSS based on fuzzy logic will now allow getting the result from a set of input parameters. All information is transmitted through HTTP requests.
A demonstration application was created to test the developed web-based application.
The demo with a graphical interface contains a working area - a two-dimensional space, which randomly places 10 circles, each of which has a randomly defined starting speed and direction of movement (Fig. 9).
Circles are all the same mass and can interact with each other and with area borders by elastic collisions.
All circles are in a closed environment and are acting due to the rules of physics. However, to test the web-based application, its functionality as a DSS, based on fuzzy logic, as well as the API, a single circle was chosen - it can change its parameters in real-time under the influence of external factors. This circle is filled with red color.
The simulating demo application checks whether there are other circles within radius N and if so, then the closest point to the subject is taken into account. If this circle is moving towards the defined one, then these two circles are considered and the following input parameters are calculated: velocity difference, angle difference, and distance. These three input parameters are crisp input for DSS and are sent in the HTTP request to the web-based application. Output crisp parameter is velocity offset for the defined red circle.
Thus, the speed of one of the circles can dynamically change and the decision of it is taken by the outer system, based on the inner system's parameters.
This should minimize the number of possible collisions if an expert human has correctly adjusted all the parameters and rules of the decision support system.
Ten tests were conducted, each lasting a minute of the internal time of the demonstration system, which does not include time for calculation, data transfer, and render.
Half of them - five tests, were conducted without the use of DSS and all circles moved according to the rules of physics. The other half of the tests were performed using DSS.
The result of each independent test is the number of collisions the red circle made with other circles during the specified internal testing system time.
Results are presented in Figure 10.
The red line specifies the number of average collisions along with five tests without the usage of DSS. The blue line specifies the average number of collisions while performing another five tests with the usage of DSS.
One of the main tasks in the nearest future is creation of a web interface for the developed application. It will definitely improve the quality of work of the human expert with the application because communication is currently supported only via HTTP requests.
Another important task is to increase the reliability of the application and provide correct encapsulation of independent data, as it allows the possibility of supporting many dependent cyberphysical systems simultaneously.
One more issue that needs to be solved, is to create an authorization for human experts, other applications that can participate in the configuration of the DSS, as well as for cyber-physical systems that are end-users by themselves.
And, finally, increasing the productivity of fuzzy logic services and improving the algorithm of processing crisp data with the fuzzy-based approach is always relevant, so that is one more task to be done in the future.
The proposed application can be used by any CPS that has access to the communication environment and has software and hardware to work with the HTTP protocol.
Usage of the API of the proposed web-based application for modeling and deploying fuzzy logicbased decision support systems for autonomous cyber-physical systems, allows creation of a desktop or a web application with a human interface so that an expert can create and configure a DSS, based on fuzzy logic.
The API can also be used by other software to automatically fill datasets and configure the DSS.
Among the disadvantages is that this web-based application cannot be used by CPSs that need to make numerous decisions in a short period of time to function properly, such as real-time systems.
There is also the question of the reliability of such a server solution, as a disruption of the remote server, hosting the web application, or network interruptions, can lead to dependent CPSs failing.